Directory UMM :Data Elmu:jurnal:UVW:World Development:Vol28.Issue11.Nov2000:
World Development Vol. 28, No. 11, pp. 1907±1925, 2000
Ó 2000 Published by Elsevier Science Ltd.
Printed in Great Britain
0305-750X/00/$ - see front matter
www.elsevier.com/locate/worlddev
PII: S0305-750X(00)00061-9
Impediments and Innovation in International
Rivers: The Waters of South Asia
BEN CROW and NIRVIKAR SINGH *
University of California, Santa Cruz, USA
Summary. Ð International cooperation over the major rivers in South Asia has recently become
much closer. Five agreements, signed in 1996 and 1997 against a background of greater regional
economic and nongovernmental contact, could facilitate signi®cant progress to mitigate ¯ooding
and drought, to provide a basis for greater regional cooperation, and to sustain irrigation
expansion and industrial development. This paper identi®es past impediments to cooperation. It
examines how new agreements seem to oer negotiation on a wider range of issues than previously,
and to expand the range of potential negotiating bodies beyond national governments to include
subnational governments, private corporations, and nongovernmental organizations. Ó 2000
Published by Elsevier Science Ltd.
Key words Ð rivers, negotiations, South Asia, water, con¯ict, cooperation
1. THE CONFLICTS AND THE
POSSIBILITIES OF SOUTH ASIAN
RIVERS
The great rivers of South Asia, particularly
the Ganges and Brahmaputra, have been the
subject of at least four decades of discussion
between governments of the region. While those
discussions have continued, until 1996 with little
productive outcome, the rivers have contributed, through ¯ood and drought, to the uncertainty and impoverishment of the lives of the
largest concentration of poor people anywhere
in the world. 1 There is nevertheless a growing
consensus that the perils of the rivers can be
turned into prosperity. This paper explores
some of the possibilities opened up by recent
innovations in international cooperation.
In this ®rst section, we describe the promise
of South Asian rivers, provide an overview of
the region's international relations over water,
and outline the new directions opened by the
agreements of 1996 and 1997. Section 2
explores the course of past diplomacy with a
case study of the Ganges river dispute between
India and Bangladesh, including how this
dispute colored a subsequent unsuccessful
attempt at regional cooperation. Section 2
concludes with a description of visions of
regional water and power development which
have been expressed by the governments of
India and Bangladesh. Section 3 introduces a
range of conceptual issues relevant for negotiations over water development: con¯ict over
the allocation of property rights, who is included in the bargaining process, the scope of their
negotiations, and the rules that govern the
process. These issues are related to the historical
* We are grateful to two referees whose painstaking and
detailed comments on an earlier version led to substantial improvement in this paper. The ®rst author thanks
Leo Rose and Geo Romm, with whom he collaborated
on a related research project, funded by the United
Nations Development Program; Shankar Bajpai, Greg
Browder, Dipak Giyawali, Ramaswamy Iyer, Obaidullah Khan, Peter Lyden, Ainun Nishat,Keith Pitman,
Zahir Sadeque, and Uggyen Tsering. We are also
grateful to Peter Kriz for helpful discussions and
comments. Financial support was received from the
University of California at Santa Cruz through the
Academic Senate; Center for Global, International and
Regional Studies; and Division of Social Sciences. The
second author also received generous support from the
University of CaliforniaÕs Institute on Global Con¯ict
and Cooperation, and from the Center for Research on
Economic Development and Policy Reform, Stanford
University, where he was a Visiting Scholar. An early
draft of this paper was published by the Henry Stimson
Institute, Washington, DC. Final revision accepted: 25
February 2000.
1907
1908
WORLD DEVELOPMENT
problems in South Asian river development
discussed in Section 2. Section 4 examines the
innovations incorporated in the ®ve treaties
signed in 1996 and 1997, and how they address
some of the past obstacles to successful agreement as discussed in Sections 2 and 3. This
section also considers directions in which
current innovations might be extended as bases
of regional cooperation. Section 5 provides a
summary conclusion.
(a) The problems and the promise of
South Asian rivers
South Asian governments seek to control the
great rivers of their region 2 because they oer
partial, but tangible, solutions to the most
fundamental problems of rural poverty, industrial constraints, and urban stress that those
governments seek to address. At present, the
ways in which control has been soughtÐ
through national visions, covert appropriation
and bilateral bargainingÐconstrain what can
be achieved.
There is a growing community of scholars,
ocials and politicians 3 in South Asia that
believes that the region's rivers can be better
harnessed in support of economic development.
For example, George Verghese, a prominent
former Indian-newspaper-editor and long-time
proponent of river development, has written:
There is no reason why the immiserised population of
this resource-rich Basin should remain poor and hostage to a recurring cycle of devastating ¯ood and
drought. There is sucient indication that international funding and technical assistance will be forthcoming in ample measure if the Basin-states decide
cooperatively to harness the waters of these mighty
rivers, green the mountains and conserve ``losing
ground.'' 4
A complementary view (Boyce, 1987), agrees
about the promise of water development, and
argues that con¯ict between rich and poor
hinders the emergence of local cooperative
institutions which could employ water resources eectively. Together these arguments make
a powerful case for those types of river development which recognize the political and
economic forces shaping current con¯ict at
international and local levels. This case was
echoed in the US Congress in 1996 with a
concurrent resolution urging South Asian
governments to ``redouble their eorts to devise
development projects that could relieve the
poverty of those people living in the Ganges
and Brahmaputra River Basin and address the
critical problems of ¯ooding and drought...'' 5
(b) Failure of past negotiations
There has been little regional cooperation in
South Asia, least of all about the contentious
topic of water. The South Asian Association
for Regional Cooperation (SAARC), established in the 1980s, provides a forum for
discussion of the least controversial topics.
However, the most heated ones, particularly
water resource negotiations, were excluded
from its brief at the start. With the exception of
one meeting in 1986, negotiations over water
have been exclusively bilateral, that is, involving only two states. India, in fact, has repeatedly insisted on this bilateralism, a point we will
take up again in Section 2.
The most heated and long-running river
disagreement has been between Bangladesh
(and its predecessor, East Pakistan) and India
over the sharing of the ¯ow of the Ganges. This
question has sometimes been temporarily
settled by interim agreements, and has occasionally erupted into internationally publicized
disagreement. More typically, as for the decade
up to 1996, it has been marked by chronic lack
of agreement: intergovernmental negotiations
of varying frequency that repeatedly fail to
make substantive progress.
The governments of India and Nepal have
had many rounds of sometimes tense negotiations relating to hydroelectricity generation,
irrigation water, and ¯ood control, and early
agreements about shared projects have been
controversial in Nepal. Water has the potential
to be NepalÕs major economic resource, and
successive governments have expected that the
sale of hydroelectric power to India would
generate signi®cant revenues for economic
development. Until 1996, little progress had
been made toward this goal.
In Section 2, two of the most prominent
elements obstructing international cooperation
will be identi®ed and described: the Indian
governmentÕs insistence on bilateral rather than
multilateral negotiations (termed bilateralism)
and competing national visions for water
development.
(c) New directions
Though these obstructions persist, recent
agreements open new directions in regional
cooperation, including:
IMPEDIMENTS AND INNOVATION IN INTERNATIONAL RIVERS
(i) shifting some negotiations from the diplomatic or governmental sphere at least partly
into the sphere of the private economy;
(ii) bringing third parties, other than governments, into negotiation, design and implementation of cooperative projects;
(iii) moving toward the sharing of eventual
bene®ts and costs, rather than establishing
®xed payments based on anticipated outcomes;
(iv) making tentative steps toward limited
multilateral discussion.
These innovations bring new ideas, processes
and parties to the negotiations and help to
address uncertainties about the natural world,
and about the outcome of human intervention
in the natural world. We brie¯y examine the
directions that these innovations suggest.
The ®rst new direction concerns the borders
of diplomacy. Social practices with distant
origins dictate that certain issues are economic
and others diplomatic. In the diplomatic
sphere, a range of issues of national interest,
possibly including economic matters, are
negotiated by governments. In the private
economic sphere various exchanges of trade,
investment, lending and labor are negotiated
between private parties, and only the broad
framework is regulated by government. The
distinction between the two spheres, and
international practices to be followed within
them, are periodically a matter of public
debate, as in the free trade debates of 18th and
19th century England, and the late 20th century
return to those debates, around various GATT
(General Agreement for Taris and Trade)
negotiations, NAFTA (North American Free
Trade Agreement) and so on. These recent
trade negotiations are signi®cant examples of
``economic diplomacy.'' It is also true that the
boundary between public and private shifts
over time. Developing country governments
such as India's intervened heavily in the
domestic economic sphere, and this carried
over to negotiations with other countries. For
example, India's ties with the Soviet Union
were represented by a complex mix of political
and economic agreements. This approach did
not, however, extend to diplomacy within
South Asia.
There are contrasts between diplomatic and
private economic practices that suggest that
there are advantages to the transfer of some
international river negotiations from diplomacy
to commerce. In the private economic sphere,
enterprises enter negotiations with clear private
1909
incentives, that is, to generate a return for
owners or shareholders. By contrast, diplomacy
involves negotiation between governments
having multiple objectives and less direct
incentives, including the approval of bureaucratic superiors and the various processes of
collective representation or protest. This set of
contrasts suggests that negotiations within the
private economic sphere have the advantages of
simple goals, clear rules and pressures for quick
completion. The shift from diplomatic to
private economic negotiation parallels the
widely debated processes of privatization
and liberalization, and is discussed in Sections 3
and 4.
The second new direction suggested by the
1996 agreements relates to the inclusion of third
parties such as corporations, local governments
and nongovernmental organizations in international negotiations. This may be advantageous if new social, economic and intellectual
resources are to be brought to bear upon
concerns shared across national boundaries.
When negotiations are shifted from diplomacy
to commerce third parties are necessarily
involved. A further widening can be seen,
however, in the growth of nongovernmental
networks involved in international negotiation
about environmental risks and possibilities.
The third new direction relates to the sharing
of costs and bene®ts of international environmental change. Situations of uncertainty
present a challenge to intergovernmental
cooperation. In the case of South Asia, climatic
and tectonic variations combined with the
unpredictable consequences of agriculture, land
clearance and other human interventions,
constitute a signi®cant source of uncertainty
in¯uencing international environmental negotiations. River ¯ows, sediment loads and
groundwater levels are only partially predictable. In addition, projects to harness natural
resources have uncertain bene®ts and costs. In
these conditions, the sharing of bene®ts and
costs constitutes a promising direction for
international cooperation. This does not, of
course, exclude governments from this risksharing: large-scale projects, in particular, will
require their participation, even if only as
guarantors or underwriters.
The fourth new direction, of multilateralism,
has parallels with the second: new resources are
brought to bear on problems, and unintended
negative impacts on those otherwise excluded
are avoided, leading to agreements that are
more likely to be stable in the long run. In
1910
WORLD DEVELOPMENT
addition, there is the possibility of expanding
the ``gains to trade'' by expanding the set of
bargainers. These issues are taken up in
Sections 3 and 4.
2. IMPEDIMENTS TO AGREEMENT:
BILATERALISM AND NATIONAL
VISIONS
International river negotiations frequently
take many decades before agreement can be
achieved. Water resource cooperation in the
basins of the rivers Ganges and Brahmaputra
may constitute the most complex of all international water negotiations. The combined
scale of the environmental, social and technical
issues has no equivalent anywhere else in the
world. Given the scale of these problems, and
the paucity of regional resources that can be
garnered to address them, 6 it is not surprising
that the negotiation of international cooperation should be protracted and uncertain.
Nevertheless, it is arguable that the past focus
on bilateral negotiations, and on national,
rather than regional, perspectives, have slowed
the achievement of cooperation and river
development. 7 Here we focus on India's policy
of bilateralism, and its consequences for India,
Bangladesh and Nepal in past river negotiations.
(a) Bilateralism
Bilateralism has been a consistent Indian
government prerequisite for negotiations with
its South Asian neighbors ever since Independence in 1947. Almost all negotiations about a
range of key issues, from river development to
trade and transit, have been negotiated on that
basis.
Rose (1987) identi®es bilateralism as one of
two main principles of Indian government
policy towards its neighbors, acceptance of
India as the major regional power being the
other. He describes bilateralism:
As de®ned by India, the South Asian system would
function through the greater coordination of IndiaÕs
bilateral economic relations with the other regional
states; any substantial integration of the economies
of the other states (e.g., Pakistan and Sri Lanka or
Nepal and Bangladesh) or any use of a multilateral
approach to regional economic issues (e.g., the river
systems of Nepal, Bhutan, Bangladesh and India)
should be discouraged.
This policy of bilateralism is a complex aspect
of Indian foreign policy. We argue here that it
constitutes a serious obstacle to achieving the
potential of South Asian water resource development. Two alternative perspectives on bilateralism can be identi®ed.
First, spokespersons for the Indian Ministry
of External Aairs have emphasized the additional complexity and duration implied by
multilateral negotiations. 8 In this perspective,
bilateral negotiations on speci®c bilateral
questions or projects are more expeditious than
multilateral negotiations. This argument is
plausible, but has to be tempered by the experience of delays in bilateral negotiations between the Indian government and its
neighbors.
A second perspective on bilateralism, sometimes found in political and academic discussion in Nepal and Bangladesh, is that
bilateralism allows India to dominate the
subcontinent, presumably by hindering the
formation of a ``bargaining coalition'' by
India's neighbors. This perspective may have
historical validity but gives little immediate
purchase on current questions of cooperation.
It is also unclear to what extent, and in what
ways, it actually impinges upon Indian
governmental discussions and decisions.
The emphasis on bilateral relations leads to a
particular focus on the sequence of issues that
have dominated the relations between two
governments. It has been argued that this focus
encourages the perception that river development is a ``zero sum game,'' a common obstacle
in international river discussions. 9 This
perception, that the gain of one country is
necessarily the loss of the other, gives the
negotiations a particular charge: any compromise of prior national objectives can be
portrayed as a victory for the other side.
Whether this perception is rational is another
matter, however: even bilateral situations may
involve mutual gains. The real question is
whether multilateralism might substantially
expand the gainsÐenough to overcome additional complexity or bargaining costs.
The focus on histories of bilateral relations
may also create fertile ground for the growth of
myths about the nature and possibilities of
those relations. In the case of India and
Bangladesh, perceptions of river negotiations
are deeply in¯uenced by the history and myths
of past negotiation over one project, IndiaÕs
Farakka Barrage across the Ganges. All
subsequent discussion about water between
IMPEDIMENTS AND INNOVATION IN INTERNATIONAL RIVERS
these two governments, and in their national
media, tends to be mired in the myths and
colored by the particular paths of past bilateral
relations.
The case that follows illustrates how bilateral
negotiations create obstacles to regional cooperation. It begins with negotiations between
India and Bangladesh, primarily about the
sharing and development of the Ganges, and
continues with how this history aected the
®rst multilateral discussion, among India,
Bangladesh and Nepal, about regional water
development.
(b) From con¯ict over the Ganges to claims on
Nepal's rivers
In South Asia, the con¯ict between India and
Bangladesh over the waters of the Ganges is
generally known as the Farakka dispute. This
name comes from the government of India's
construction, in the 1960s, of a large barrage 10
across the Ganges at the small town of
Farakka, 11 miles upstream from the Indian
border with Bangladesh. The barrage was
constructed to divert water down the river
Hooghly (a distributary of the Ganges) to
improve the access of ships to Calcutta Port,
which is connected by the Hooghly to the sea.
The government of Pakistan and, after 1971,
the government of Bangladesh opposed the
construction and operation of the barrage
because they perceived (rightly) that the
reduction in the dry season ¯ow of the Ganges
would have serious implications for East
Pakistan/Bangladesh.
The dispute has long since ceased to be
primarily about the water diverted by the
Farakka Barrage. These waters have been
overshadowed by larger (but never publicly
calculated) diversions further upstream for the
development of irrigated agriculture in India.
The con¯ict over the sharing of the Ganges
arises primarily because the dry season ¯ow of
that river cannot sustain the irrigation needs, as
currently estimated (possibly with exaggeration) of both countries.
The con¯ict over the Ganges can best be
understood, at least initially, by examining the
dierent chronological phases through which it
has developed. 11 Table 1 summarizes those
phases. In the 1950s, shipping companies using
the state of West Bengal's Calcutta Port, and
the industrialists and politicians of the state,
thought that the port's decline could be
reversed if water from the river Ganges
could be diverted to ``¯ush'' the channel of the
river Hooghly. The ``¯ushing'' thesis was, and
remains, controversial. 12 Nevertheless, the
government of India decided in the early 1960s
that the barrage should be constructed. Indian
Prime Minister Jawarhalal Nehru pronounced
himself satis®ed that there would be no significant consequences for East Pakistan. The
government of Pakistan occasionally objected
to the Barrage throughout its construction
during the 1960s.
When East Pakistan gained independence as
Bangladesh in 1971, the Farakka Barrage was
still under construction. Independence brought
to power the Awami (PeopleÕs) League, which
had recently made a pledge that ``[e]very
instrument of foreign policy must be immediately utilized to secure a just solution of
this [the Farakka] problem.'' 13 Relations
between Bangladesh and India were, nonetheless, much more cordial than they had been
between Pakistan and India. A 25-year ``Treaty
of Friendship'' was signed, and, the new
Bangladesh government began to shift from the
complete opposition of its predecessor to
negotiations about the operation of the
Barrage, and an interim agreement for trial
operations was signed in 1974.
This phase of the con¯ict was brought to an
end with the electoral defeat of the Congress
Party government in India in 1977, and the
signing, six months later, of the 1977 Ganges
Waters Agreement. The Agreement laid down
the division of the dry season ¯ow in the
Ganges, and began a phase of discussions
about how best to increase that ¯ow in the
lower reaches of the Ganges. The incompatibility of national visions of water development
Table 1. The four phases of the Ganges river dispute
Phase
Phase
Phase
Phase
Phase
I
II
III
IV
1911
Period
Focus
1951±71
1971±77
1977±82
1983±87
Discussion over the Farakka Barrage
Division of dry season ¯ow (leading to Ganges Waters Agreement)
Augmentation of the ¯ow
Independent national river development
1912
WORLD DEVELOPMENT
emerging in these discussions meant that no
agreement could be reached on how to increase
the dry season ¯ows. After brief extensions to
the sharing agreement, minus the written
guarantee of minimum dry season ¯ows,
discussions about ¯ow augmentation ended,
and the agreement lapsed.
The Farakka Barrage created a dispute about
the waters of the Ganges that has in¯uenced,
and continues to in¯uence, subsequent discussion of water policy. The focus on IndiaÕs
Farakka Barrage, encouraged by the bilateral
boundaries of the discussion, seems to generate
a narrow, nationalistic discussion that ignores
wider possibilities and sustains debilitating
myths. Two of these myths are as follows. In
Bangladesh, the Farakka Barrage has been
widely portrayed in political and media
discussions as a symbol of IndiaÕs evil intent
toward Bangladesh. Technical controversy
about the ``¯ushing'' process, as well as IndiaÕs
failure to recognize the downstream consequences of the project, left space for the assertion that the barrage was built because of its
deleterious eects on Bangladesh (then Pakistan). 14 A second myth is the assertion that
India can cause ¯ooding in Bangladesh through
the release of water stored behind the Farakka
Barrage. In fact ( note 10) the barrage can store
far too little to have a signi®cant eect on
¯oods in Bangladesh.
These myths, with complex foundations in
the colonial division of the subcontinent, as
well as in the technical uncertainties and
ambiguities of water development, posit negotiations over water as a ``zero-sum game.'' This
structuring of the discussion leaves little space
for the possibility that water development
could be an enterprise from which all sides gain
much more than they lose. The boundaries of
discussion could be relaxed in the context of
regional, in place of bilateral, discussion.
The next part of this case is the story of how
the governments of India and Bangladesh
arranged a meeting with the government of
Nepal to discuss river waters. The agenda of
the meeting was almost entirely determined by
the concerns of the two larger states, and the
bilateral discussions they had previously
generated. Their history of bilateral relations
ensured that the concerns of Nepal were omitted from the agenda. In the aftermath of a 1985
cyclone devastating some of the southern
islands of Bangladesh, Indian Prime Minister
Rajiv Gandhi ¯ew to Bangladesh to oer
Indian assistance. He met Bangladesh President
Ershad and, amongst other topics, they
discussed the question of sharing the Ganges.
This meeting began a new series of discussions
and initiatives and an interim ``memorandum
of understanding'' was signed in November
1985. They met again the next year, and agreed
that the two governments should make a joint
approach to the government of Nepal. Despite
considerable ambiguities about the terms of
reference of the meeting, and what was not to
be discussed, the meeting took place in Kathmandu in October of 1986.
An Indian participant describes the representatives of Nepal at this three-day meeting
``repeating Ômutual bene®tÕ like a mantra'' in
response to all requests and discussions. 15
Whatever the two leaders may have intended,
the meeting was preceded by limitations agreed
between the relevant ministers of Bangladesh
and India. Indian ministers requested and
received assurances from Bangladesh that this
meeting would not constitute a precedent for
trilateral negotiations, and letters sent by
Bangladesh and India to Nepal further limited
the meeting to seeking information and data.
The outcome was a strangely restricted meeting: two governments were making a formal
approach to discuss the development of the
rivers of a third country, having agreed in
advance that they would not allow the third
government any standing in their deliberations.
They were only going to ask the third government to supply them with the information
needed to plan development of rivers in the
third governmentÕs territory.
The understandable response from Nepal
was to ask ``why?'' What is the mutual bene®t
of this cooperation, and what is the standing of
Nepal in this development of its own resources?
The representatives of the governments of India
and Bangladesh were unable to answer this
question without prejudicing their prior agreement that the meeting should not be a precedent for trilateral negotiations. The two visiting
government delegations left empty handed:
Nepal gave none of the requested information
about water resources.
This history illustrates that nationally-constituted visions of water resource development
frequently overlook the concerns of neighbors.
Thus, when India and Bangladesh approached
Nepal, the concerns of Nepal were overlooked.
In addition, when India decided to build the
Farakka Barrage in the early 1960s, Nehru was
convinced (presumably by his engineers) that it
would cause ``no real injury'' downstream. This
IMPEDIMENTS AND INNOVATION IN INTERNATIONAL RIVERS
also is the case when governments are planning
new water development, as the next section
illustrates.
(c) National visions
Three recent visions of water development
illustrate particular national perspectives and
the process by which claims are made over
water. For many years, engineering advisors to
the governments of East Pakistan and Bangladesh argued that the shortage of dry-season
water in East Bengal could best be tackled by
the construction of water storage reservoirs in
the Himalayan headwater rivers of the Ganges,
in India and Nepal. The 1977 Ganges Waters
Agreement gave both governments a mandate
to produce a vision of how to augment the dry
season ¯ow in the lower Ganges. In a 1978
report (updated in 1983), the government of
Bangladesh presented its proposal for how this
might be done: more than 80 large water storage reservoirs could be constructed in India and
Nepal to increase the dry season ¯ow in the
Ganges. This report estimated that the release
of stored water could meet the irrigation needs
of all three countries. If constructed, the storage of melting snows from the Himalayas and
of monsoon precipitation would contribute to
the mitigation of ¯oods. The release of stored
water could also generate urgently needed
electricity for the subcontinent. This vision has
faded. The government of Nepal was sometimes willing to consider the idea: construction
of at least some of the dams could generate
revenue for Nepal. The government of India
has so far been opposed, on grounds of multilateral ineciency and engineering implausibility.
While the Indian government opposed any
Bangladesh claim to water stored on the
Ganges, the Indian government suggested that
the Ganges ¯ow could be increased by transferring water from the Brahmaputra. They
proposed (in a limited circulation report
published in 1978 and updated in 1983) (i) a
200-mile long and 900 ft wide canal passing
through northwest Bangladesh, linking the
Brahmaputra to the Ganges, (ii) a barrage
across the Brahmaputra, upstream of Bangladesh, to divert water into the canal, and (iii) (at
some later date) three large dams on the
Brahmaputra and Meghna rivers. This vision
has also faded. The experience with Farakka
provides reason for Bangladesh to avoid an
Indian-controlled barrage across the Brah-
1913
maputra. Instead, Bangladesh has quietly
investigated the possibility of internal transfers,
controlled by Bangladesh, and using less landconsuming, existing river channels rather than
a huge canal.
The third vision would give land-locked
Nepal access to the sea. Included in the 1978
proposal from Bangladesh was the idea that a
canal could be built from Nepal, across a
narrow part of India separating Nepal from
Bangladesh, to join up with a large river in
Northern Bangladesh. This idea had been
presented by the government of Bangladesh on
NepalÕs behalf. It expressed NepalÕs frustration
at having its connections to the global economy
controlled by India, and subjected to imposition of ``trade and transit'' treaties (and, at one
time, a trade embargo). It also represented an
attempt by two smaller countries to formulate
common interests against a shared experience
of dominance by the much larger, and
economically more powerful, neighbor, but was
dismissed by the Indian government.
What do we conclude from these three
visions of water development in South Asia?
Each expresses a national vision making little
accommodation to the concerns of other states.
There has been no recognition that compromise
might achieve greater bene®ts for the region.
To some extent, these visions were shaped by
the failure of conventional bilateral diplomacy.
We next explore conceptual issues with respect
to cooperation over international river waters
in South Asia, and in Section 4, we examine
recent innovations and future directions.
3. FACILITATING COOPERATION:
CONCEPTUAL ISSUES
Cooperation can occur when mutual bene®ts
are possible. The existence of mutual gains is
not, however, sucient for cooperation: the
prisoner's dilemma is the most famous example
of failure to achieve mutual gains. In the prisoner's dilemma game, the inability to communicate and to commit to a binding agreement is
the source of the problem, given the structure
of the payos. Ways around the prisoner's
dilemma involve changing the structure of
payos, through repeated play, introduction of
certain kinds of incomplete information about
the payos or player types, 16 or the introduction of other features that broaden the game.
Technically, the game remains one of noncooperative behavior, but such behavior can
1914
WORLD DEVELOPMENT
support a cooperative outcome in the expanded
game. While some types of incomplete information help resolve the prisoner's dilemma, in
other types of games, incomplete information
makes achievement of mutual bene®ts limited
or even impossible, this being a version of the
``lemons problem.'' 17
In game theory, the cooperative approach to
negotiating for mutual bene®t assumes that
binding agreements can be signed, but typically
relies on normative axioms to identify a cooperative solution, such as the Nash bargaining
solution, without examining the details of the
bargaining: protocols, rules and so on.
Noncooperative game theoretic approaches try
to model these details. Recent theoretical
advances have begun to connect the noncooperative and cooperative approaches to negotiation or bargaining, by identifying which
noncooperative bargaining protocols will lead
to standard axiomatic cooperative outcomes
such as the Nash bargaining solution. These
theories have been extended from the original
two-player analysis to cover multiple bargaining parties.
The abstract discussion above helps to focus
on the following questions in discussing how
cooperation in the use of water resources may be
facilitated. First, are there truly potential mutual
bene®ts, or is the situation one of con¯ict (one
party can only gain at another's expense)?
Second, if the answer to the ®rst question is
negative, can the situation be rede®ned (e.g.,
broadened in scope by considering other
dimensions, or repeated interaction) to transform it to one of potential mutual bene®t? Third,
what are the impediments to actually achieving
mutual bene®ts: uncertainty, asymmetries of
information, exclusion of key parties aected by
the transaction, or inecient bargaining protocols? Obviously, the answers to all three questions will overlap. We will explore the experience
of international negotiations over the use of
South Asia's water resources in this framework.
To illustrate, we brie¯y consider the Indus
waters case, alluded to earlier.
Who ``owned'' the ®ve rivers making up the
Indus system was the basis of dispute between
India and Pakistan after independence and
partition in 1947. There was uncertainty about
property rights, which made any de facto
property rights of limited value. But, any division of the ¯ows was viewed as potentially
providing a gain to one side at the expense of
the other. This is the common problem in the
allocation of property rights: mutual bene®ts
from agreement are unclear or nonexistent. The
Indus Waters Treaty of 1960 established that
the rivers would be divided between India and
Pakistan. This division resolved the property
rights, but to either Pakistan or IndiaÕs disadvantage, depending on who bore the cost of
investments required to make the proposed
division workable. Intervention that changed
the nature of the game came in the form of
external funding via the World Bank. Secure
property rights and ®nancing from the World
Bank allowed each country's share of the
system to provide much more than could have
been achieved while property rights were
uncertain.
The example illustrates answers to the ®rst
two questions above: a situation of con¯ict was
transformed in a simple way to one of mutual
bene®t, by a third party ``sweetening the deal.''
The answers, and the solution to the problem,
however, are both limited. The theoretical issue
of why the World Bank intervened can be
answered brie¯y: the geopolitical incentives
underlying that institution and its backers
determined this role. The practical solution to
the Indus waters problem was limited, however,
because the waters that ``belong'' to Pakistan
partly ¯ow through Indian territory. If optimal
usage of these waters requires large ®xed
investments, a dierent set of issues, well
beyond the allocation of property rights to
water, arise. In fact, these issues have prevented
such investment occurring: analysis of the
problems brings one back to the examples
discussed in the previous section. We turn to
these issues after discussing the problem of
allocation of property rights and other
conceptual issues.
(a) Con¯ict and cooperation
A situation of pure con¯ict is one where no
mutual bene®ts are possible: in a bilateral
negotiation, a gain for one party must result in
a loss for the other. A simple reallocation of
property rights is therefore a situation of pure
con¯ict. Therefore, to the extent that international river disputes are disputes over property
rights, one would be pessimistic about resolving
con¯ict. Only when property rights are sorted
out can mutually bene®cial agreements
contingent on those rights be contemplated.
Several factors soften this pessimism, and
provide the basis for our subsequent analysis.
While the geography of rivers and underground
aquifers creates de facto property rights, even
IMPEDIMENTS AND INNOVATION IN INTERNATIONAL RIVERS
when there is no explicit agreement on rights,
these property rights may be uncertain enough
that a certain right to less may be more valuable than an uncertain claim to more. Thus the
removal of uncertainty on one or both sides in
a bilateral negotiation may create the scope for
mutually bene®cial agreement. One example of
the bene®ts of the removal of uncertainty
would be in the perceived returns to investments that support the eective use of water, or
that require reasonably certain water supplies.
A slightly dierent way of looking at the
possibility of mutual bene®t in a con¯ict over
property rights is as follows. De facto rights
may be subject to overt con¯ict or threats,
which may be resolved through agreement. To
put it in technical terms, the disagreement point
in the bargaining game may itself be endogenous, and this endogenous location of the
disagreement point creates the possibility of
mutual bene®ts from agreement. For example,
one may identify this as a possible factor in the
Indus waters dispute, where both sides wished
to avoid a broader con¯ict that would have
altered the disagreement point.
A third way of going beyond a situation of
pure con¯ict is to expand the dimension of the
bargaining space. The relinquishing of a property right, or a claim, can be compensated by a
transfer in the opposite direction, just as in any
mutually bene®cial trade. The transfer in this
case may be of money, material goods or
intangibles such as security. This need not
involve going outside the sphere of conventional diplomacy (including commercial diplomacy), but private exchanges may be helpful,
for reasons we discuss below.
While multilateral negotiations do not
provide any direct advantage with respect to
the resolution of property rights con¯icts, the
dierent aspects of ameliorating con¯ict that
we have outlined apply to multilateral as well as
bilateral bargaining situations.
(b) Property rights and investment
While we have noted above the bene®ts for
water-related investment of a transformation of
uncertain claims into certain property rights,
we earlier pointed out the limits of this transformation in the Indus case. Here we explain
these limits. As we remarked earlier, the river
waters allocated to Pakistan through the Indus
treaty partly ¯ow through Indian territory.
Clearly, consumptive uses of this waterÐsuch
as irrigationÐby India are ruled out by the
1915
treaty. What about nonconsumptive uses such
as hydroelectric power? In principle, India
should be able to negotiate such uses and
undertake the investments required. Pakistan
might, however, desire to monitor the investment to make sure that no water is being
diverted. Such monitoring would raise
currently insurmountable issues of security and
sovereignty. Even greater problems would arise
with respect to Pakistani investment within
Indian territory. As a result of these issues, the
full hydroelectric potential of the Chenab and
Jhelum remains untapped. 18
(c) Monetization vs. barter
Bilateral barter in international diplomacy,
as elsewhere, is subject to the need to ®nd a
``double coincidence of wants.'' Monetization,
in this context, is the establishment of mutually
agreed-upon values to services, enabling more
general exchange to proceed within the
complex fabric of diplomatic relations. Besides
the extreme case of overcoming the lack of a
double coincidence of wants, monetization
more generally expands the set of gains from
trade, since transfers of money more fully
transfer value than does barter.
(d) Bilateralism vs. multilateralism
As we have detailed, bilateralism has been an
important aspect of India's policy with respect
to its South Asian neighbors. Bilateralism may
be justi®ed for all parties in terms of simplicity
of negotiations or, only for some, as a way of
avoiding opposing coalitions and preserving
bargaining power. But, in the case of rivers
¯owing through more than two countries, or
where an entire river basin spans more than
two territories, bilateral bargaining may neglect
positive and negative externalities, and limit the
mutual bene®ts of possible agreements on
water development and usage. Bilateralism
combined with conventional barter diplomacy
may also limit the gains from trade, though in
this case it is the lack of fully transferable value
that is the culprit, rather than bilateralism.
Hence it is the existence of externalities that
provides a case for multilateralism over bilateralism in river negotiations.
(e) Private exchange vs. diplomacy
Conventional diplomacy is characterized by
barter, either involving speci®c items, or of
1916
WORLD DEVELOPMENT
broader scope (general reciprocity). More and
more, over time, diplomacy extends to the
commercial sphere, covering international trade
and investment in particular. Again, this can
involve speci®c exchanges between governments, or instead the setting of rules under which
private parties operate. Whether the actual
exchange takes place between governments or
private parties depends crucially on who owns
the potential objects of trade.
Natural resources such as river waters have
conventionally been treated as governmentowned, and therefore international negotiations over their shared development and use
have been ®rmly in the sphere of diplomacy,
albeit with economic components and
economic implications. Where private parties,
such as farmers and industrialists, have been
the ultimate users of water (for irrigation,
power and navigation), they have had only an
indirect say in such negotiations, through
political in¯uence or political pressure. This
political model has also governed the domestic
allocation of water at subnational levels,
through politically-determined pricing and
investment subsidies.
What changes with private involvement in
water development decisions, whether at the
subnational or the international level? To the
extent that ownership is transferred to private
entities, decision-making will be determined by
dierent objective functions. Private entities
may range from corporations that maximize
pro®ts to nongovernmental organizations
(NGOs) that maximize some aggregate
measure of their members' welfare. Though
governments may also theoretically maximize
aggregate welfare, in practice, the incentive
mechanisms to enforce this may be too weak,
particularly at the national level.
Even if ownership is not privatized, when
private parties are involved in sharing the costs
and bene®ts of water development, their
objectives will have a more direct impact on
decision-making than in the conventional
model of political in¯uence. We can think of
their participation in contracting and bargaining as similar to multilateralism in extending
the set of those who bargain. If this helps to
internalize externalities, then greater eciency
in bargaining may be realized. A further bene®t
goes beyond overcoming externalities. While
the countries in a multilateral negotiation are
determined by geography, private entities such
as multinational ®rms can be asked to compete
for seats at the bargaining table, enhancing the
potential gains to others involved in the negotiation.
The inclusion of private parties in negotiation
over water development and use not only
changes objectives, but makes them generally
more transparent. National governments may
not have easily identi®able objectives, since they
are a complex mix of the preferences of
constituents, politicians (the agents of constituents) and bureaucrats (the agents of politicians). Lower level governments provide some
degree of disaggregation, but subnational
private entities are required to be more open
about their goals and performance than is
traditional for governments in South Asia. One
can conjecture that greater transparency will, on
the whole, aid agreement in water negotiations.
There are two ®nal implications of the
inclusion of private parties. First, there is,
perhaps, greater ¯exibility in the kinds of
contracts that can be signed. In principle, there
is nothing to stop governments from signing
commercial contracts (including those specifying sharing of costs and bene®ts), but there may
be problems due to incomplete information: in
particular, the appearance of possible impropriety may prevent even the signing of honest
contracts. This assumes that pro®t-making
entities have better internal monitoring and
control mechanisms. A stronger argument is
based on commitment. Sovereign governments
may not be able to commit credibly to certain
kinds of agreements, while private parties can.
This simply re¯ects the nature of sovereignty.
Governments can expropriate and renege on
contracts involving private parties, but this may
involve greater reputation loss than breaking or
bending vaguely worded treaties.
Overall, therefore, it may be seen that the
inclusion of private or nongovernmental entities in negotiations over water development and
use implies changes more profound than those
involved in shifting to multilateralism or to
monetization. At the same time, the role of
private parties would be impossible or severely
limited without both those changes. We turn
now to the recent experience in South Asia, to
explore how the above factors may play out in
the future.
4. INNOVATIONS AT THE
INTERNATIONAL LEVEL
When representatives of the Nepalese government, in the trilateral meeting of October 1986,
IMPEDIMENTS AND INNOVATION IN INTERNATIONAL RIVERS
asked the governments of India and Bangladesh to spell out the ``mutual bene®ts'' of joint
river development, they were asking two
fundamental and related questions. How will
Nepal bene®t from the water development
proposed by India and Bangladesh? What can
be achieved through regional cooperation that
could not be achieved bilaterally? There is at
least one answer to the latter question: Table 2
lists what governments have sought from each
other in relation to regional water resources.
The realization of a signi®cant proportion of
these, or similar, proposals could constitute an
important advance, mitigating power and irrigation-water shortages, ¯ood and drought, and
contributing to the potential for regional
stability and economic development.
The realization of these transactions will,
however, be limited under a strictly bilateral
regime because the possible multiple uses of
river water and the basin-wide consequences of
many development proposals do not respect
territorial boundaries. For example, the development of hydroelectric power bilaterally by
Nepal and India may preclude consideration of
other potential water storage services. Flood
mitigation and irrigation expansion, for
Bangladesh and eastern India's plains states,
could be achieved through the construction of
water storage in Nepal. If that construction is
1917
undertaken bilaterally by Nepal and India, it is
less likely that services other than hydroelectric
power generation will be given serious consideration, although, innovations in water use and
negotiations over water development are
beginning to relax past limitations.
By examining innovations at the international level, we aim to bring out the general
principles that can transform international
negotiations over water rights and usage. These
general principles include rights allocation
mechanisms, governing institutions, and rules
for exchange.
(a) Five international agreements
1996 was a great year for river treaties in
South Asia. In January, Nepal and India signed
a treaty advancing a decades-old river development proposal, and in its last month India
and Bangladesh signed a 30-year treaty seeking
to resolve the dispute between the two countries
over the sharing of the Ganges waters. In
between, two other agreements were signed,
establishing procedures for power supply from
Nepal and Bhutan to India. A ®fth, more
tentative agreement came in April 1997, when
representatives of India, Bhutan, Nepal and
Bangladesh considered forming a subregional
economic group within the SAARC framework
Table 2. Potential international transactions in South Asiaa
Potential parties
Nepal to India
India to Nepal
India to Bangladesh
Bangladesh to India
Bangladesh to Nepal
Nepal to Bangladesh
Bhutan to India
India to Bhutan
a
Good or service
Type of exchange
anticipated
Supply of hydro-electric powerd
Supply of water storage bene®tsb;d
Navigation and transitc;d
Provision of ®nance for constructiond
Provision of engineering expertised
Supply of water storage bene®tsd
Granting secure expectations of minimum ¯owd
Navigation and transit rightsd
Transfer of water from Brahmaputra to Gangese
Navigation and transit rightse
Supply of hydro-electric powere
Supply of water storage bene®tse
Supply of hydro-electric powerc
Supply of water storage bene®tse
Navigation and transitc
Provision of ®nance and engineering for
constructionc
Monetized
Barter exchange
Barter exchange
Monetized
Probably monetized
Barter exchange
Barter exchange
Barter exchange
Barter exchange
Barter exchange
Monetized
Barter exchange
Monetized
Barter exchange
Barter exchange
Partly monetized
Source: Crow et al. (1995, Chapter 8, Table 18).
Including water storage for dry season irr
Ó 2000 Published by Elsevier Science Ltd.
Printed in Great Britain
0305-750X/00/$ - see front matter
www.elsevier.com/locate/worlddev
PII: S0305-750X(00)00061-9
Impediments and Innovation in International
Rivers: The Waters of South Asia
BEN CROW and NIRVIKAR SINGH *
University of California, Santa Cruz, USA
Summary. Ð International cooperation over the major rivers in South Asia has recently become
much closer. Five agreements, signed in 1996 and 1997 against a background of greater regional
economic and nongovernmental contact, could facilitate signi®cant progress to mitigate ¯ooding
and drought, to provide a basis for greater regional cooperation, and to sustain irrigation
expansion and industrial development. This paper identi®es past impediments to cooperation. It
examines how new agreements seem to oer negotiation on a wider range of issues than previously,
and to expand the range of potential negotiating bodies beyond national governments to include
subnational governments, private corporations, and nongovernmental organizations. Ó 2000
Published by Elsevier Science Ltd.
Key words Ð rivers, negotiations, South Asia, water, con¯ict, cooperation
1. THE CONFLICTS AND THE
POSSIBILITIES OF SOUTH ASIAN
RIVERS
The great rivers of South Asia, particularly
the Ganges and Brahmaputra, have been the
subject of at least four decades of discussion
between governments of the region. While those
discussions have continued, until 1996 with little
productive outcome, the rivers have contributed, through ¯ood and drought, to the uncertainty and impoverishment of the lives of the
largest concentration of poor people anywhere
in the world. 1 There is nevertheless a growing
consensus that the perils of the rivers can be
turned into prosperity. This paper explores
some of the possibilities opened up by recent
innovations in international cooperation.
In this ®rst section, we describe the promise
of South Asian rivers, provide an overview of
the region's international relations over water,
and outline the new directions opened by the
agreements of 1996 and 1997. Section 2
explores the course of past diplomacy with a
case study of the Ganges river dispute between
India and Bangladesh, including how this
dispute colored a subsequent unsuccessful
attempt at regional cooperation. Section 2
concludes with a description of visions of
regional water and power development which
have been expressed by the governments of
India and Bangladesh. Section 3 introduces a
range of conceptual issues relevant for negotiations over water development: con¯ict over
the allocation of property rights, who is included in the bargaining process, the scope of their
negotiations, and the rules that govern the
process. These issues are related to the historical
* We are grateful to two referees whose painstaking and
detailed comments on an earlier version led to substantial improvement in this paper. The ®rst author thanks
Leo Rose and Geo Romm, with whom he collaborated
on a related research project, funded by the United
Nations Development Program; Shankar Bajpai, Greg
Browder, Dipak Giyawali, Ramaswamy Iyer, Obaidullah Khan, Peter Lyden, Ainun Nishat,Keith Pitman,
Zahir Sadeque, and Uggyen Tsering. We are also
grateful to Peter Kriz for helpful discussions and
comments. Financial support was received from the
University of California at Santa Cruz through the
Academic Senate; Center for Global, International and
Regional Studies; and Division of Social Sciences. The
second author also received generous support from the
University of CaliforniaÕs Institute on Global Con¯ict
and Cooperation, and from the Center for Research on
Economic Development and Policy Reform, Stanford
University, where he was a Visiting Scholar. An early
draft of this paper was published by the Henry Stimson
Institute, Washington, DC. Final revision accepted: 25
February 2000.
1907
1908
WORLD DEVELOPMENT
problems in South Asian river development
discussed in Section 2. Section 4 examines the
innovations incorporated in the ®ve treaties
signed in 1996 and 1997, and how they address
some of the past obstacles to successful agreement as discussed in Sections 2 and 3. This
section also considers directions in which
current innovations might be extended as bases
of regional cooperation. Section 5 provides a
summary conclusion.
(a) The problems and the promise of
South Asian rivers
South Asian governments seek to control the
great rivers of their region 2 because they oer
partial, but tangible, solutions to the most
fundamental problems of rural poverty, industrial constraints, and urban stress that those
governments seek to address. At present, the
ways in which control has been soughtÐ
through national visions, covert appropriation
and bilateral bargainingÐconstrain what can
be achieved.
There is a growing community of scholars,
ocials and politicians 3 in South Asia that
believes that the region's rivers can be better
harnessed in support of economic development.
For example, George Verghese, a prominent
former Indian-newspaper-editor and long-time
proponent of river development, has written:
There is no reason why the immiserised population of
this resource-rich Basin should remain poor and hostage to a recurring cycle of devastating ¯ood and
drought. There is sucient indication that international funding and technical assistance will be forthcoming in ample measure if the Basin-states decide
cooperatively to harness the waters of these mighty
rivers, green the mountains and conserve ``losing
ground.'' 4
A complementary view (Boyce, 1987), agrees
about the promise of water development, and
argues that con¯ict between rich and poor
hinders the emergence of local cooperative
institutions which could employ water resources eectively. Together these arguments make
a powerful case for those types of river development which recognize the political and
economic forces shaping current con¯ict at
international and local levels. This case was
echoed in the US Congress in 1996 with a
concurrent resolution urging South Asian
governments to ``redouble their eorts to devise
development projects that could relieve the
poverty of those people living in the Ganges
and Brahmaputra River Basin and address the
critical problems of ¯ooding and drought...'' 5
(b) Failure of past negotiations
There has been little regional cooperation in
South Asia, least of all about the contentious
topic of water. The South Asian Association
for Regional Cooperation (SAARC), established in the 1980s, provides a forum for
discussion of the least controversial topics.
However, the most heated ones, particularly
water resource negotiations, were excluded
from its brief at the start. With the exception of
one meeting in 1986, negotiations over water
have been exclusively bilateral, that is, involving only two states. India, in fact, has repeatedly insisted on this bilateralism, a point we will
take up again in Section 2.
The most heated and long-running river
disagreement has been between Bangladesh
(and its predecessor, East Pakistan) and India
over the sharing of the ¯ow of the Ganges. This
question has sometimes been temporarily
settled by interim agreements, and has occasionally erupted into internationally publicized
disagreement. More typically, as for the decade
up to 1996, it has been marked by chronic lack
of agreement: intergovernmental negotiations
of varying frequency that repeatedly fail to
make substantive progress.
The governments of India and Nepal have
had many rounds of sometimes tense negotiations relating to hydroelectricity generation,
irrigation water, and ¯ood control, and early
agreements about shared projects have been
controversial in Nepal. Water has the potential
to be NepalÕs major economic resource, and
successive governments have expected that the
sale of hydroelectric power to India would
generate signi®cant revenues for economic
development. Until 1996, little progress had
been made toward this goal.
In Section 2, two of the most prominent
elements obstructing international cooperation
will be identi®ed and described: the Indian
governmentÕs insistence on bilateral rather than
multilateral negotiations (termed bilateralism)
and competing national visions for water
development.
(c) New directions
Though these obstructions persist, recent
agreements open new directions in regional
cooperation, including:
IMPEDIMENTS AND INNOVATION IN INTERNATIONAL RIVERS
(i) shifting some negotiations from the diplomatic or governmental sphere at least partly
into the sphere of the private economy;
(ii) bringing third parties, other than governments, into negotiation, design and implementation of cooperative projects;
(iii) moving toward the sharing of eventual
bene®ts and costs, rather than establishing
®xed payments based on anticipated outcomes;
(iv) making tentative steps toward limited
multilateral discussion.
These innovations bring new ideas, processes
and parties to the negotiations and help to
address uncertainties about the natural world,
and about the outcome of human intervention
in the natural world. We brie¯y examine the
directions that these innovations suggest.
The ®rst new direction concerns the borders
of diplomacy. Social practices with distant
origins dictate that certain issues are economic
and others diplomatic. In the diplomatic
sphere, a range of issues of national interest,
possibly including economic matters, are
negotiated by governments. In the private
economic sphere various exchanges of trade,
investment, lending and labor are negotiated
between private parties, and only the broad
framework is regulated by government. The
distinction between the two spheres, and
international practices to be followed within
them, are periodically a matter of public
debate, as in the free trade debates of 18th and
19th century England, and the late 20th century
return to those debates, around various GATT
(General Agreement for Taris and Trade)
negotiations, NAFTA (North American Free
Trade Agreement) and so on. These recent
trade negotiations are signi®cant examples of
``economic diplomacy.'' It is also true that the
boundary between public and private shifts
over time. Developing country governments
such as India's intervened heavily in the
domestic economic sphere, and this carried
over to negotiations with other countries. For
example, India's ties with the Soviet Union
were represented by a complex mix of political
and economic agreements. This approach did
not, however, extend to diplomacy within
South Asia.
There are contrasts between diplomatic and
private economic practices that suggest that
there are advantages to the transfer of some
international river negotiations from diplomacy
to commerce. In the private economic sphere,
enterprises enter negotiations with clear private
1909
incentives, that is, to generate a return for
owners or shareholders. By contrast, diplomacy
involves negotiation between governments
having multiple objectives and less direct
incentives, including the approval of bureaucratic superiors and the various processes of
collective representation or protest. This set of
contrasts suggests that negotiations within the
private economic sphere have the advantages of
simple goals, clear rules and pressures for quick
completion. The shift from diplomatic to
private economic negotiation parallels the
widely debated processes of privatization
and liberalization, and is discussed in Sections 3
and 4.
The second new direction suggested by the
1996 agreements relates to the inclusion of third
parties such as corporations, local governments
and nongovernmental organizations in international negotiations. This may be advantageous if new social, economic and intellectual
resources are to be brought to bear upon
concerns shared across national boundaries.
When negotiations are shifted from diplomacy
to commerce third parties are necessarily
involved. A further widening can be seen,
however, in the growth of nongovernmental
networks involved in international negotiation
about environmental risks and possibilities.
The third new direction relates to the sharing
of costs and bene®ts of international environmental change. Situations of uncertainty
present a challenge to intergovernmental
cooperation. In the case of South Asia, climatic
and tectonic variations combined with the
unpredictable consequences of agriculture, land
clearance and other human interventions,
constitute a signi®cant source of uncertainty
in¯uencing international environmental negotiations. River ¯ows, sediment loads and
groundwater levels are only partially predictable. In addition, projects to harness natural
resources have uncertain bene®ts and costs. In
these conditions, the sharing of bene®ts and
costs constitutes a promising direction for
international cooperation. This does not, of
course, exclude governments from this risksharing: large-scale projects, in particular, will
require their participation, even if only as
guarantors or underwriters.
The fourth new direction, of multilateralism,
has parallels with the second: new resources are
brought to bear on problems, and unintended
negative impacts on those otherwise excluded
are avoided, leading to agreements that are
more likely to be stable in the long run. In
1910
WORLD DEVELOPMENT
addition, there is the possibility of expanding
the ``gains to trade'' by expanding the set of
bargainers. These issues are taken up in
Sections 3 and 4.
2. IMPEDIMENTS TO AGREEMENT:
BILATERALISM AND NATIONAL
VISIONS
International river negotiations frequently
take many decades before agreement can be
achieved. Water resource cooperation in the
basins of the rivers Ganges and Brahmaputra
may constitute the most complex of all international water negotiations. The combined
scale of the environmental, social and technical
issues has no equivalent anywhere else in the
world. Given the scale of these problems, and
the paucity of regional resources that can be
garnered to address them, 6 it is not surprising
that the negotiation of international cooperation should be protracted and uncertain.
Nevertheless, it is arguable that the past focus
on bilateral negotiations, and on national,
rather than regional, perspectives, have slowed
the achievement of cooperation and river
development. 7 Here we focus on India's policy
of bilateralism, and its consequences for India,
Bangladesh and Nepal in past river negotiations.
(a) Bilateralism
Bilateralism has been a consistent Indian
government prerequisite for negotiations with
its South Asian neighbors ever since Independence in 1947. Almost all negotiations about a
range of key issues, from river development to
trade and transit, have been negotiated on that
basis.
Rose (1987) identi®es bilateralism as one of
two main principles of Indian government
policy towards its neighbors, acceptance of
India as the major regional power being the
other. He describes bilateralism:
As de®ned by India, the South Asian system would
function through the greater coordination of IndiaÕs
bilateral economic relations with the other regional
states; any substantial integration of the economies
of the other states (e.g., Pakistan and Sri Lanka or
Nepal and Bangladesh) or any use of a multilateral
approach to regional economic issues (e.g., the river
systems of Nepal, Bhutan, Bangladesh and India)
should be discouraged.
This policy of bilateralism is a complex aspect
of Indian foreign policy. We argue here that it
constitutes a serious obstacle to achieving the
potential of South Asian water resource development. Two alternative perspectives on bilateralism can be identi®ed.
First, spokespersons for the Indian Ministry
of External Aairs have emphasized the additional complexity and duration implied by
multilateral negotiations. 8 In this perspective,
bilateral negotiations on speci®c bilateral
questions or projects are more expeditious than
multilateral negotiations. This argument is
plausible, but has to be tempered by the experience of delays in bilateral negotiations between the Indian government and its
neighbors.
A second perspective on bilateralism, sometimes found in political and academic discussion in Nepal and Bangladesh, is that
bilateralism allows India to dominate the
subcontinent, presumably by hindering the
formation of a ``bargaining coalition'' by
India's neighbors. This perspective may have
historical validity but gives little immediate
purchase on current questions of cooperation.
It is also unclear to what extent, and in what
ways, it actually impinges upon Indian
governmental discussions and decisions.
The emphasis on bilateral relations leads to a
particular focus on the sequence of issues that
have dominated the relations between two
governments. It has been argued that this focus
encourages the perception that river development is a ``zero sum game,'' a common obstacle
in international river discussions. 9 This
perception, that the gain of one country is
necessarily the loss of the other, gives the
negotiations a particular charge: any compromise of prior national objectives can be
portrayed as a victory for the other side.
Whether this perception is rational is another
matter, however: even bilateral situations may
involve mutual gains. The real question is
whether multilateralism might substantially
expand the gainsÐenough to overcome additional complexity or bargaining costs.
The focus on histories of bilateral relations
may also create fertile ground for the growth of
myths about the nature and possibilities of
those relations. In the case of India and
Bangladesh, perceptions of river negotiations
are deeply in¯uenced by the history and myths
of past negotiation over one project, IndiaÕs
Farakka Barrage across the Ganges. All
subsequent discussion about water between
IMPEDIMENTS AND INNOVATION IN INTERNATIONAL RIVERS
these two governments, and in their national
media, tends to be mired in the myths and
colored by the particular paths of past bilateral
relations.
The case that follows illustrates how bilateral
negotiations create obstacles to regional cooperation. It begins with negotiations between
India and Bangladesh, primarily about the
sharing and development of the Ganges, and
continues with how this history aected the
®rst multilateral discussion, among India,
Bangladesh and Nepal, about regional water
development.
(b) From con¯ict over the Ganges to claims on
Nepal's rivers
In South Asia, the con¯ict between India and
Bangladesh over the waters of the Ganges is
generally known as the Farakka dispute. This
name comes from the government of India's
construction, in the 1960s, of a large barrage 10
across the Ganges at the small town of
Farakka, 11 miles upstream from the Indian
border with Bangladesh. The barrage was
constructed to divert water down the river
Hooghly (a distributary of the Ganges) to
improve the access of ships to Calcutta Port,
which is connected by the Hooghly to the sea.
The government of Pakistan and, after 1971,
the government of Bangladesh opposed the
construction and operation of the barrage
because they perceived (rightly) that the
reduction in the dry season ¯ow of the Ganges
would have serious implications for East
Pakistan/Bangladesh.
The dispute has long since ceased to be
primarily about the water diverted by the
Farakka Barrage. These waters have been
overshadowed by larger (but never publicly
calculated) diversions further upstream for the
development of irrigated agriculture in India.
The con¯ict over the sharing of the Ganges
arises primarily because the dry season ¯ow of
that river cannot sustain the irrigation needs, as
currently estimated (possibly with exaggeration) of both countries.
The con¯ict over the Ganges can best be
understood, at least initially, by examining the
dierent chronological phases through which it
has developed. 11 Table 1 summarizes those
phases. In the 1950s, shipping companies using
the state of West Bengal's Calcutta Port, and
the industrialists and politicians of the state,
thought that the port's decline could be
reversed if water from the river Ganges
could be diverted to ``¯ush'' the channel of the
river Hooghly. The ``¯ushing'' thesis was, and
remains, controversial. 12 Nevertheless, the
government of India decided in the early 1960s
that the barrage should be constructed. Indian
Prime Minister Jawarhalal Nehru pronounced
himself satis®ed that there would be no significant consequences for East Pakistan. The
government of Pakistan occasionally objected
to the Barrage throughout its construction
during the 1960s.
When East Pakistan gained independence as
Bangladesh in 1971, the Farakka Barrage was
still under construction. Independence brought
to power the Awami (PeopleÕs) League, which
had recently made a pledge that ``[e]very
instrument of foreign policy must be immediately utilized to secure a just solution of
this [the Farakka] problem.'' 13 Relations
between Bangladesh and India were, nonetheless, much more cordial than they had been
between Pakistan and India. A 25-year ``Treaty
of Friendship'' was signed, and, the new
Bangladesh government began to shift from the
complete opposition of its predecessor to
negotiations about the operation of the
Barrage, and an interim agreement for trial
operations was signed in 1974.
This phase of the con¯ict was brought to an
end with the electoral defeat of the Congress
Party government in India in 1977, and the
signing, six months later, of the 1977 Ganges
Waters Agreement. The Agreement laid down
the division of the dry season ¯ow in the
Ganges, and began a phase of discussions
about how best to increase that ¯ow in the
lower reaches of the Ganges. The incompatibility of national visions of water development
Table 1. The four phases of the Ganges river dispute
Phase
Phase
Phase
Phase
Phase
I
II
III
IV
1911
Period
Focus
1951±71
1971±77
1977±82
1983±87
Discussion over the Farakka Barrage
Division of dry season ¯ow (leading to Ganges Waters Agreement)
Augmentation of the ¯ow
Independent national river development
1912
WORLD DEVELOPMENT
emerging in these discussions meant that no
agreement could be reached on how to increase
the dry season ¯ows. After brief extensions to
the sharing agreement, minus the written
guarantee of minimum dry season ¯ows,
discussions about ¯ow augmentation ended,
and the agreement lapsed.
The Farakka Barrage created a dispute about
the waters of the Ganges that has in¯uenced,
and continues to in¯uence, subsequent discussion of water policy. The focus on IndiaÕs
Farakka Barrage, encouraged by the bilateral
boundaries of the discussion, seems to generate
a narrow, nationalistic discussion that ignores
wider possibilities and sustains debilitating
myths. Two of these myths are as follows. In
Bangladesh, the Farakka Barrage has been
widely portrayed in political and media
discussions as a symbol of IndiaÕs evil intent
toward Bangladesh. Technical controversy
about the ``¯ushing'' process, as well as IndiaÕs
failure to recognize the downstream consequences of the project, left space for the assertion that the barrage was built because of its
deleterious eects on Bangladesh (then Pakistan). 14 A second myth is the assertion that
India can cause ¯ooding in Bangladesh through
the release of water stored behind the Farakka
Barrage. In fact ( note 10) the barrage can store
far too little to have a signi®cant eect on
¯oods in Bangladesh.
These myths, with complex foundations in
the colonial division of the subcontinent, as
well as in the technical uncertainties and
ambiguities of water development, posit negotiations over water as a ``zero-sum game.'' This
structuring of the discussion leaves little space
for the possibility that water development
could be an enterprise from which all sides gain
much more than they lose. The boundaries of
discussion could be relaxed in the context of
regional, in place of bilateral, discussion.
The next part of this case is the story of how
the governments of India and Bangladesh
arranged a meeting with the government of
Nepal to discuss river waters. The agenda of
the meeting was almost entirely determined by
the concerns of the two larger states, and the
bilateral discussions they had previously
generated. Their history of bilateral relations
ensured that the concerns of Nepal were omitted from the agenda. In the aftermath of a 1985
cyclone devastating some of the southern
islands of Bangladesh, Indian Prime Minister
Rajiv Gandhi ¯ew to Bangladesh to oer
Indian assistance. He met Bangladesh President
Ershad and, amongst other topics, they
discussed the question of sharing the Ganges.
This meeting began a new series of discussions
and initiatives and an interim ``memorandum
of understanding'' was signed in November
1985. They met again the next year, and agreed
that the two governments should make a joint
approach to the government of Nepal. Despite
considerable ambiguities about the terms of
reference of the meeting, and what was not to
be discussed, the meeting took place in Kathmandu in October of 1986.
An Indian participant describes the representatives of Nepal at this three-day meeting
``repeating Ômutual bene®tÕ like a mantra'' in
response to all requests and discussions. 15
Whatever the two leaders may have intended,
the meeting was preceded by limitations agreed
between the relevant ministers of Bangladesh
and India. Indian ministers requested and
received assurances from Bangladesh that this
meeting would not constitute a precedent for
trilateral negotiations, and letters sent by
Bangladesh and India to Nepal further limited
the meeting to seeking information and data.
The outcome was a strangely restricted meeting: two governments were making a formal
approach to discuss the development of the
rivers of a third country, having agreed in
advance that they would not allow the third
government any standing in their deliberations.
They were only going to ask the third government to supply them with the information
needed to plan development of rivers in the
third governmentÕs territory.
The understandable response from Nepal
was to ask ``why?'' What is the mutual bene®t
of this cooperation, and what is the standing of
Nepal in this development of its own resources?
The representatives of the governments of India
and Bangladesh were unable to answer this
question without prejudicing their prior agreement that the meeting should not be a precedent for trilateral negotiations. The two visiting
government delegations left empty handed:
Nepal gave none of the requested information
about water resources.
This history illustrates that nationally-constituted visions of water resource development
frequently overlook the concerns of neighbors.
Thus, when India and Bangladesh approached
Nepal, the concerns of Nepal were overlooked.
In addition, when India decided to build the
Farakka Barrage in the early 1960s, Nehru was
convinced (presumably by his engineers) that it
would cause ``no real injury'' downstream. This
IMPEDIMENTS AND INNOVATION IN INTERNATIONAL RIVERS
also is the case when governments are planning
new water development, as the next section
illustrates.
(c) National visions
Three recent visions of water development
illustrate particular national perspectives and
the process by which claims are made over
water. For many years, engineering advisors to
the governments of East Pakistan and Bangladesh argued that the shortage of dry-season
water in East Bengal could best be tackled by
the construction of water storage reservoirs in
the Himalayan headwater rivers of the Ganges,
in India and Nepal. The 1977 Ganges Waters
Agreement gave both governments a mandate
to produce a vision of how to augment the dry
season ¯ow in the lower Ganges. In a 1978
report (updated in 1983), the government of
Bangladesh presented its proposal for how this
might be done: more than 80 large water storage reservoirs could be constructed in India and
Nepal to increase the dry season ¯ow in the
Ganges. This report estimated that the release
of stored water could meet the irrigation needs
of all three countries. If constructed, the storage of melting snows from the Himalayas and
of monsoon precipitation would contribute to
the mitigation of ¯oods. The release of stored
water could also generate urgently needed
electricity for the subcontinent. This vision has
faded. The government of Nepal was sometimes willing to consider the idea: construction
of at least some of the dams could generate
revenue for Nepal. The government of India
has so far been opposed, on grounds of multilateral ineciency and engineering implausibility.
While the Indian government opposed any
Bangladesh claim to water stored on the
Ganges, the Indian government suggested that
the Ganges ¯ow could be increased by transferring water from the Brahmaputra. They
proposed (in a limited circulation report
published in 1978 and updated in 1983) (i) a
200-mile long and 900 ft wide canal passing
through northwest Bangladesh, linking the
Brahmaputra to the Ganges, (ii) a barrage
across the Brahmaputra, upstream of Bangladesh, to divert water into the canal, and (iii) (at
some later date) three large dams on the
Brahmaputra and Meghna rivers. This vision
has also faded. The experience with Farakka
provides reason for Bangladesh to avoid an
Indian-controlled barrage across the Brah-
1913
maputra. Instead, Bangladesh has quietly
investigated the possibility of internal transfers,
controlled by Bangladesh, and using less landconsuming, existing river channels rather than
a huge canal.
The third vision would give land-locked
Nepal access to the sea. Included in the 1978
proposal from Bangladesh was the idea that a
canal could be built from Nepal, across a
narrow part of India separating Nepal from
Bangladesh, to join up with a large river in
Northern Bangladesh. This idea had been
presented by the government of Bangladesh on
NepalÕs behalf. It expressed NepalÕs frustration
at having its connections to the global economy
controlled by India, and subjected to imposition of ``trade and transit'' treaties (and, at one
time, a trade embargo). It also represented an
attempt by two smaller countries to formulate
common interests against a shared experience
of dominance by the much larger, and
economically more powerful, neighbor, but was
dismissed by the Indian government.
What do we conclude from these three
visions of water development in South Asia?
Each expresses a national vision making little
accommodation to the concerns of other states.
There has been no recognition that compromise
might achieve greater bene®ts for the region.
To some extent, these visions were shaped by
the failure of conventional bilateral diplomacy.
We next explore conceptual issues with respect
to cooperation over international river waters
in South Asia, and in Section 4, we examine
recent innovations and future directions.
3. FACILITATING COOPERATION:
CONCEPTUAL ISSUES
Cooperation can occur when mutual bene®ts
are possible. The existence of mutual gains is
not, however, sucient for cooperation: the
prisoner's dilemma is the most famous example
of failure to achieve mutual gains. In the prisoner's dilemma game, the inability to communicate and to commit to a binding agreement is
the source of the problem, given the structure
of the payos. Ways around the prisoner's
dilemma involve changing the structure of
payos, through repeated play, introduction of
certain kinds of incomplete information about
the payos or player types, 16 or the introduction of other features that broaden the game.
Technically, the game remains one of noncooperative behavior, but such behavior can
1914
WORLD DEVELOPMENT
support a cooperative outcome in the expanded
game. While some types of incomplete information help resolve the prisoner's dilemma, in
other types of games, incomplete information
makes achievement of mutual bene®ts limited
or even impossible, this being a version of the
``lemons problem.'' 17
In game theory, the cooperative approach to
negotiating for mutual bene®t assumes that
binding agreements can be signed, but typically
relies on normative axioms to identify a cooperative solution, such as the Nash bargaining
solution, without examining the details of the
bargaining: protocols, rules and so on.
Noncooperative game theoretic approaches try
to model these details. Recent theoretical
advances have begun to connect the noncooperative and cooperative approaches to negotiation or bargaining, by identifying which
noncooperative bargaining protocols will lead
to standard axiomatic cooperative outcomes
such as the Nash bargaining solution. These
theories have been extended from the original
two-player analysis to cover multiple bargaining parties.
The abstract discussion above helps to focus
on the following questions in discussing how
cooperation in the use of water resources may be
facilitated. First, are there truly potential mutual
bene®ts, or is the situation one of con¯ict (one
party can only gain at another's expense)?
Second, if the answer to the ®rst question is
negative, can the situation be rede®ned (e.g.,
broadened in scope by considering other
dimensions, or repeated interaction) to transform it to one of potential mutual bene®t? Third,
what are the impediments to actually achieving
mutual bene®ts: uncertainty, asymmetries of
information, exclusion of key parties aected by
the transaction, or inecient bargaining protocols? Obviously, the answers to all three questions will overlap. We will explore the experience
of international negotiations over the use of
South Asia's water resources in this framework.
To illustrate, we brie¯y consider the Indus
waters case, alluded to earlier.
Who ``owned'' the ®ve rivers making up the
Indus system was the basis of dispute between
India and Pakistan after independence and
partition in 1947. There was uncertainty about
property rights, which made any de facto
property rights of limited value. But, any division of the ¯ows was viewed as potentially
providing a gain to one side at the expense of
the other. This is the common problem in the
allocation of property rights: mutual bene®ts
from agreement are unclear or nonexistent. The
Indus Waters Treaty of 1960 established that
the rivers would be divided between India and
Pakistan. This division resolved the property
rights, but to either Pakistan or IndiaÕs disadvantage, depending on who bore the cost of
investments required to make the proposed
division workable. Intervention that changed
the nature of the game came in the form of
external funding via the World Bank. Secure
property rights and ®nancing from the World
Bank allowed each country's share of the
system to provide much more than could have
been achieved while property rights were
uncertain.
The example illustrates answers to the ®rst
two questions above: a situation of con¯ict was
transformed in a simple way to one of mutual
bene®t, by a third party ``sweetening the deal.''
The answers, and the solution to the problem,
however, are both limited. The theoretical issue
of why the World Bank intervened can be
answered brie¯y: the geopolitical incentives
underlying that institution and its backers
determined this role. The practical solution to
the Indus waters problem was limited, however,
because the waters that ``belong'' to Pakistan
partly ¯ow through Indian territory. If optimal
usage of these waters requires large ®xed
investments, a dierent set of issues, well
beyond the allocation of property rights to
water, arise. In fact, these issues have prevented
such investment occurring: analysis of the
problems brings one back to the examples
discussed in the previous section. We turn to
these issues after discussing the problem of
allocation of property rights and other
conceptual issues.
(a) Con¯ict and cooperation
A situation of pure con¯ict is one where no
mutual bene®ts are possible: in a bilateral
negotiation, a gain for one party must result in
a loss for the other. A simple reallocation of
property rights is therefore a situation of pure
con¯ict. Therefore, to the extent that international river disputes are disputes over property
rights, one would be pessimistic about resolving
con¯ict. Only when property rights are sorted
out can mutually bene®cial agreements
contingent on those rights be contemplated.
Several factors soften this pessimism, and
provide the basis for our subsequent analysis.
While the geography of rivers and underground
aquifers creates de facto property rights, even
IMPEDIMENTS AND INNOVATION IN INTERNATIONAL RIVERS
when there is no explicit agreement on rights,
these property rights may be uncertain enough
that a certain right to less may be more valuable than an uncertain claim to more. Thus the
removal of uncertainty on one or both sides in
a bilateral negotiation may create the scope for
mutually bene®cial agreement. One example of
the bene®ts of the removal of uncertainty
would be in the perceived returns to investments that support the eective use of water, or
that require reasonably certain water supplies.
A slightly dierent way of looking at the
possibility of mutual bene®t in a con¯ict over
property rights is as follows. De facto rights
may be subject to overt con¯ict or threats,
which may be resolved through agreement. To
put it in technical terms, the disagreement point
in the bargaining game may itself be endogenous, and this endogenous location of the
disagreement point creates the possibility of
mutual bene®ts from agreement. For example,
one may identify this as a possible factor in the
Indus waters dispute, where both sides wished
to avoid a broader con¯ict that would have
altered the disagreement point.
A third way of going beyond a situation of
pure con¯ict is to expand the dimension of the
bargaining space. The relinquishing of a property right, or a claim, can be compensated by a
transfer in the opposite direction, just as in any
mutually bene®cial trade. The transfer in this
case may be of money, material goods or
intangibles such as security. This need not
involve going outside the sphere of conventional diplomacy (including commercial diplomacy), but private exchanges may be helpful,
for reasons we discuss below.
While multilateral negotiations do not
provide any direct advantage with respect to
the resolution of property rights con¯icts, the
dierent aspects of ameliorating con¯ict that
we have outlined apply to multilateral as well as
bilateral bargaining situations.
(b) Property rights and investment
While we have noted above the bene®ts for
water-related investment of a transformation of
uncertain claims into certain property rights,
we earlier pointed out the limits of this transformation in the Indus case. Here we explain
these limits. As we remarked earlier, the river
waters allocated to Pakistan through the Indus
treaty partly ¯ow through Indian territory.
Clearly, consumptive uses of this waterÐsuch
as irrigationÐby India are ruled out by the
1915
treaty. What about nonconsumptive uses such
as hydroelectric power? In principle, India
should be able to negotiate such uses and
undertake the investments required. Pakistan
might, however, desire to monitor the investment to make sure that no water is being
diverted. Such monitoring would raise
currently insurmountable issues of security and
sovereignty. Even greater problems would arise
with respect to Pakistani investment within
Indian territory. As a result of these issues, the
full hydroelectric potential of the Chenab and
Jhelum remains untapped. 18
(c) Monetization vs. barter
Bilateral barter in international diplomacy,
as elsewhere, is subject to the need to ®nd a
``double coincidence of wants.'' Monetization,
in this context, is the establishment of mutually
agreed-upon values to services, enabling more
general exchange to proceed within the
complex fabric of diplomatic relations. Besides
the extreme case of overcoming the lack of a
double coincidence of wants, monetization
more generally expands the set of gains from
trade, since transfers of money more fully
transfer value than does barter.
(d) Bilateralism vs. multilateralism
As we have detailed, bilateralism has been an
important aspect of India's policy with respect
to its South Asian neighbors. Bilateralism may
be justi®ed for all parties in terms of simplicity
of negotiations or, only for some, as a way of
avoiding opposing coalitions and preserving
bargaining power. But, in the case of rivers
¯owing through more than two countries, or
where an entire river basin spans more than
two territories, bilateral bargaining may neglect
positive and negative externalities, and limit the
mutual bene®ts of possible agreements on
water development and usage. Bilateralism
combined with conventional barter diplomacy
may also limit the gains from trade, though in
this case it is the lack of fully transferable value
that is the culprit, rather than bilateralism.
Hence it is the existence of externalities that
provides a case for multilateralism over bilateralism in river negotiations.
(e) Private exchange vs. diplomacy
Conventional diplomacy is characterized by
barter, either involving speci®c items, or of
1916
WORLD DEVELOPMENT
broader scope (general reciprocity). More and
more, over time, diplomacy extends to the
commercial sphere, covering international trade
and investment in particular. Again, this can
involve speci®c exchanges between governments, or instead the setting of rules under which
private parties operate. Whether the actual
exchange takes place between governments or
private parties depends crucially on who owns
the potential objects of trade.
Natural resources such as river waters have
conventionally been treated as governmentowned, and therefore international negotiations over their shared development and use
have been ®rmly in the sphere of diplomacy,
albeit with economic components and
economic implications. Where private parties,
such as farmers and industrialists, have been
the ultimate users of water (for irrigation,
power and navigation), they have had only an
indirect say in such negotiations, through
political in¯uence or political pressure. This
political model has also governed the domestic
allocation of water at subnational levels,
through politically-determined pricing and
investment subsidies.
What changes with private involvement in
water development decisions, whether at the
subnational or the international level? To the
extent that ownership is transferred to private
entities, decision-making will be determined by
dierent objective functions. Private entities
may range from corporations that maximize
pro®ts to nongovernmental organizations
(NGOs) that maximize some aggregate
measure of their members' welfare. Though
governments may also theoretically maximize
aggregate welfare, in practice, the incentive
mechanisms to enforce this may be too weak,
particularly at the national level.
Even if ownership is not privatized, when
private parties are involved in sharing the costs
and bene®ts of water development, their
objectives will have a more direct impact on
decision-making than in the conventional
model of political in¯uence. We can think of
their participation in contracting and bargaining as similar to multilateralism in extending
the set of those who bargain. If this helps to
internalize externalities, then greater eciency
in bargaining may be realized. A further bene®t
goes beyond overcoming externalities. While
the countries in a multilateral negotiation are
determined by geography, private entities such
as multinational ®rms can be asked to compete
for seats at the bargaining table, enhancing the
potential gains to others involved in the negotiation.
The inclusion of private parties in negotiation
over water development and use not only
changes objectives, but makes them generally
more transparent. National governments may
not have easily identi®able objectives, since they
are a complex mix of the preferences of
constituents, politicians (the agents of constituents) and bureaucrats (the agents of politicians). Lower level governments provide some
degree of disaggregation, but subnational
private entities are required to be more open
about their goals and performance than is
traditional for governments in South Asia. One
can conjecture that greater transparency will, on
the whole, aid agreement in water negotiations.
There are two ®nal implications of the
inclusion of private parties. First, there is,
perhaps, greater ¯exibility in the kinds of
contracts that can be signed. In principle, there
is nothing to stop governments from signing
commercial contracts (including those specifying sharing of costs and bene®ts), but there may
be problems due to incomplete information: in
particular, the appearance of possible impropriety may prevent even the signing of honest
contracts. This assumes that pro®t-making
entities have better internal monitoring and
control mechanisms. A stronger argument is
based on commitment. Sovereign governments
may not be able to commit credibly to certain
kinds of agreements, while private parties can.
This simply re¯ects the nature of sovereignty.
Governments can expropriate and renege on
contracts involving private parties, but this may
involve greater reputation loss than breaking or
bending vaguely worded treaties.
Overall, therefore, it may be seen that the
inclusion of private or nongovernmental entities in negotiations over water development and
use implies changes more profound than those
involved in shifting to multilateralism or to
monetization. At the same time, the role of
private parties would be impossible or severely
limited without both those changes. We turn
now to the recent experience in South Asia, to
explore how the above factors may play out in
the future.
4. INNOVATIONS AT THE
INTERNATIONAL LEVEL
When representatives of the Nepalese government, in the trilateral meeting of October 1986,
IMPEDIMENTS AND INNOVATION IN INTERNATIONAL RIVERS
asked the governments of India and Bangladesh to spell out the ``mutual bene®ts'' of joint
river development, they were asking two
fundamental and related questions. How will
Nepal bene®t from the water development
proposed by India and Bangladesh? What can
be achieved through regional cooperation that
could not be achieved bilaterally? There is at
least one answer to the latter question: Table 2
lists what governments have sought from each
other in relation to regional water resources.
The realization of a signi®cant proportion of
these, or similar, proposals could constitute an
important advance, mitigating power and irrigation-water shortages, ¯ood and drought, and
contributing to the potential for regional
stability and economic development.
The realization of these transactions will,
however, be limited under a strictly bilateral
regime because the possible multiple uses of
river water and the basin-wide consequences of
many development proposals do not respect
territorial boundaries. For example, the development of hydroelectric power bilaterally by
Nepal and India may preclude consideration of
other potential water storage services. Flood
mitigation and irrigation expansion, for
Bangladesh and eastern India's plains states,
could be achieved through the construction of
water storage in Nepal. If that construction is
1917
undertaken bilaterally by Nepal and India, it is
less likely that services other than hydroelectric
power generation will be given serious consideration, although, innovations in water use and
negotiations over water development are
beginning to relax past limitations.
By examining innovations at the international level, we aim to bring out the general
principles that can transform international
negotiations over water rights and usage. These
general principles include rights allocation
mechanisms, governing institutions, and rules
for exchange.
(a) Five international agreements
1996 was a great year for river treaties in
South Asia. In January, Nepal and India signed
a treaty advancing a decades-old river development proposal, and in its last month India
and Bangladesh signed a 30-year treaty seeking
to resolve the dispute between the two countries
over the sharing of the Ganges waters. In
between, two other agreements were signed,
establishing procedures for power supply from
Nepal and Bhutan to India. A ®fth, more
tentative agreement came in April 1997, when
representatives of India, Bhutan, Nepal and
Bangladesh considered forming a subregional
economic group within the SAARC framework
Table 2. Potential international transactions in South Asiaa
Potential parties
Nepal to India
India to Nepal
India to Bangladesh
Bangladesh to India
Bangladesh to Nepal
Nepal to Bangladesh
Bhutan to India
India to Bhutan
a
Good or service
Type of exchange
anticipated
Supply of hydro-electric powerd
Supply of water storage bene®tsb;d
Navigation and transitc;d
Provision of ®nance for constructiond
Provision of engineering expertised
Supply of water storage bene®tsd
Granting secure expectations of minimum ¯owd
Navigation and transit rightsd
Transfer of water from Brahmaputra to Gangese
Navigation and transit rightse
Supply of hydro-electric powere
Supply of water storage bene®tse
Supply of hydro-electric powerc
Supply of water storage bene®tse
Navigation and transitc
Provision of ®nance and engineering for
constructionc
Monetized
Barter exchange
Barter exchange
Monetized
Probably monetized
Barter exchange
Barter exchange
Barter exchange
Barter exchange
Barter exchange
Monetized
Barter exchange
Monetized
Barter exchange
Barter exchange
Partly monetized
Source: Crow et al. (1995, Chapter 8, Table 18).
Including water storage for dry season irr